A pathway that stays sensitised
The claim is that a synapse can be left more sensitive after brief, intense use, and that this change lasts far longer than the stimulation itself. Terje Lømo first observed it in 1966, running electrical stimulation into the perforant pathway of an anaesthetised rabbit and recording the response in the dentate gyrus, a part of the hippocampus. A short burst of high-frequency stimulation left the recorded response larger than before, and that enlargement persisted well beyond the stimulation. Timothy Bliss, who joined the Oslo laboratory in 1968, worked with Lømo to characterise the effect properly, and their 1973 paper gave the first full account of what would later be named long-term potentiation, now treated as the leading cellular candidate for how the brain stores what it learns.
Naming it, formally
The original work used anaesthetised rabbits and direct electrical stimulation of a specific hippocampal circuit, the perforant pathway feeding into the dentate gyrus, with electrodes recording the size of the postsynaptic response before and after a brief tetanic burst. That electrophysiological method, stimulate and record, remains the basic approach for studying the effect, extended over the following decades to hippocampal slices kept alive outside the animal and to genetically modified mice. Molecular work built on top of that method traced an early, short-lived phase driven by calcium entering postsynaptic cells through NMDA receptors and existing receptors being made more responsive, and a later phase, requiring new protein production, that involves physical growth in the connecting structures between neurons.
A mechanism for Hebb’s rule
What has held up is the basic phenomenon and its dependence, in many though not all pathways, on the NMDA receptor. Blocking that receptor with the drug APV, as Richard Morris did in 1986, stops potentiation from being produced in hippocampal tissue and, in the same animals, stops them learning to navigate a water maze, tying the cellular effect to a measurable behaviour rather than leaving it as a laboratory curiosity. Later experiments extended the case: mice engineered to carry an extra receptor subunit showed both enhanced potentiation and better spatial learning, and rats trained on an avoidance task showed the same receptor changes in their synapses as rats given the artificial stimulation directly, with previously trained synapses resistant to further potentiation in a way that mirrors the laboratory effect closely.
From synapse to receptor
What has not held up as cleanly is any single, complete molecular account of how the change is maintained over time. A protein called PKMζ was for years treated as essential to keeping the late phase of potentiation going, but mice bred without it still showed normal potentiation, undercutting a mechanism that had looked settled. Whether the presynaptic side of the connection, not just the postsynaptic side, contributes to the lasting change is still disputed, and researchers disagree about what signal, if any, would need to travel backward across the synapse to produce it. Where in the cell the new proteins needed for the lasting phase are actually made, near the synapse itself or back in the cell body, is also unresolved, meaning several of the field’s mechanistic claims remain provisional rather than settled.
Blocking it, and watching learning fail
The effect matters beyond the hippocampal slice because it gives researchers a physical target for conditions defined, until now, almost entirely by behaviour. Amyloid-beta, the protein fragment implicated in Alzheimer’s disease, has been shown to interfere with potentiation in the hippocampus, offering one candidate explanation for the memory decline that marks the disease early on. Addiction researchers have started treating compulsive drug use as a powerful, unwanted form of the same learning process, pointing to potentiation-like changes in the reward circuitry of the brain rather than in the hippocampus. Neither line of work amounts to a treatment yet, but both depend on the idea, established by Bliss and Lømo’s original result, that learning and memory have a traceable cellular signature rather than being purely a matter of psychology.
Necessary, not sufficient
Yes, though it rewards patience rather than delivering a single dramatic payoff. What makes it worth the time is watching a simple electrophysiological observation in an anaesthetised rabbit grow, over five decades, into a framework that touches memory, addiction and neurodegenerative disease, while never quite closing the gap between a stronger synapse and an actual stored memory. Readers who want certainty should know going in that some of the field’s earlier landmark explanations, the role of one particular maintenance protein among them, have since been walked back rather than confirmed. That openness is part of what makes the subject credible: this is a case where the caveats are as informative as the headline result, and worth reading past.